Age-Based Methods to Explore Time-Related Variables in Occupational Epidemiology Studies
Although age is recognized as the strongest predictor of mortality in chronic disease epidemiology, a calendar-based approach is often employed when evaluating time-related variables. An age-based analysis file, created by determining the value of each time-dependent variable for each age that a cohort member is followed, provides a clear definition of age at exposure and allows development of diverse analytic models. To demonstrate methods, the relationship between cancer mortality and external radiation was analyzed with Poisson regression for 14,095 Oak Ridge National Laboratory workers. Based on previous analysis of this cohort, a model with ten-year lagged cumulative radiation doses partitioned by receipt before (dose-young) or after (dose-old) age 45 was examined. Dose-response estimates were similar to calendar-year-based results with elevated risk for dose-old, but not when film badge readings were weekly before 1957. Complementary results showed increasing risk with older hire ages and earlier birth cohorts, since workers hired after age 45 were born before 1915, and dose-young and dose-old were distributed differently by birth cohorts. Risks were generally higher for smokingrelated than non-smoking-related cancers. It was difficult to single out specific variables associated with elevated cancer mortality because of: (1) birth cohort differences in hire age and mortality experience completeness, and (2) time-period differences in working conditions, dose potential, and exposure assessment. This research demonstrated the utility and versatility of the age-based approach.
- Research Organization:
- Oak Ridge Institute for Science and Education (ORISE), Oak Ridge, TN
- Sponsoring Organization:
- National Institute for Occupational Safety and Health
- DOE Contract Number:
- DE-AC05-06OR23100
- OSTI ID:
- 901098
- Report Number(s):
- orise-07-oewh-0234; TRN: US200713%%128
- Resource Relation:
- Conference: 2005 Joint Statistical Meetings, American Statistical Association Section on Statistics in Epidemiology, Alexandria, Va., August 7-11, 2005
- Country of Publication:
- United States
- Language:
- English
Similar Records
Comparison of All Solid Cancer Mortality and Incidence Dose-Response in the Life Span Study of Atomic Bomb Survivors, 1958–2009
A mortality study of workers at seven beryllium processing plants
Related Subjects
DISEASES
EPIDEMIOLOGY
MORTALITY
NEOPLASMS
ORNL
PHOTOGRAPHIC FILM DOSEMETERS
RADIATION DOSES
RADIATIONS
STATISTICS
WORKING CONDITIONS
statistics
epidemiology
occupational epidemiology
time-related variables
radiation
age at exposure